{"version":1,"generator":"agentprose Cloud docs","generated":"2026-10-18T00:00:00.000Z","site":{"name":"agentprose Cloud","description":"Documentation for agentprose Cloud: make a content site ready for AI agents, and see which agents came, what they asked and what they quoted.","url":"https://agentprose.cloud/","language":"en","feeds":["https://agentprose.cloud/docs/feed.xml"]},"posts":[{"url":"https://agentprose.cloud/docs/getting-started","slug":"getting-started","title":"Getting started","summary":"Set up agentprose Cloud on your site in about five minutes: create a site, paste one script tag, and watch AI agents use your content.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["setup","quickstart"],"categories":["Start here"],"lang":"en","readingMinutes":2,"text":"agentprose Cloud makes a blog, docs site or any content site easy for AI agents to use, and shows you what those agents do there. It has two parts that switch on together: - Agent-ready content. The open-source agentprose script gives AI browsers proper tools on every page: search your posts, read one cleanly, check a quote before citing you. See How AI agents use your site. - Your AI audience. Every tool call an agent makes shows up in your dashboard within seconds: which agent came, what it searched for, what it read and what it couldn't find. See Agent analytics. 1. Create an account and a site Sign up (free, no card), confirm your email with the code we send, then click Add site . Give it a name and the address of your site, for example https://blog.example.com/ . The domain list says which sites may send data with this site's key; it starts with the domain of the address you gave. Each site gets a public key ( pk … ). It isn't a secret: it only works from the domains on the site's list. 2. Paste one line into your site's header Use your site's real key from the site's Settings & install tab. Where to paste it on WordPress, Ghost, Webflow, Hugo and others is in Install the script. 3. Let us index your content When you add a site we crawl it to build its hosted search index. We look for agentprose.json , then your RSS/Atom feed, then your sitemap. If the site's Search index tab says no posts found , read How we find your content: usually the fix is a feed or one line of st","md":"/docs/getting-started.md"},{"url":"https://agentprose.cloud/docs/install","slug":"install","title":"Install the script","summary":"Where to paste the agentprose Cloud script tag on WordPress, Ghost, Webflow, static site generators and any other site, plus CSP, CDN and analytics-only options.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["setup","install","csp"],"categories":["Start here"],"lang":"en","readingMinutes":1,"text":"The whole install is one script tag in the <head of every page: It replaces any agentprose tag you already have. agentprose Cloud serves the open-source agentprose library itself, so your site doesn't depend on a public CDN, and the script knows where its Cloud is. https://agentprose.cloud/v/0.1.0/agentprose.min.js never changes and is cached for a year; https://agentprose.cloud/agentprose.min.js always points at the current version. The script is deferred, about 9 KB gzipped, and doesn't touch your layout. It makes no network requests on page load. It only loads its search engine when an agent first calls a site-wide tool. Where to paste it Platform Where --- --- WordPress Your theme's header settings, or any \"insert headers and footers\" plugin Ghost Settings → Code injection → Site header Webflow Site settings → Custom code → Head code Hugo, Jekyll, Astro, Eleventy, Next.js Your base layout's <head Anything else Wherever your pages' <head is built Check that it works Open a page with ?agentprose-inspect added to the address. The 🔌 inspector lists the registered tools and lets you call them. Content-Security-Policy With a strict Content-Security-Policy, allow https://agentprose.cloud in script-src (the script loads its search engine and adapters from the same folder) and in connect-src (beacons and hosted search). The script uses no eval and no inline code. Load it from jsDelivr instead The same library is on npm. From a public CDN the script can't know your Cloud, so data-","md":"/docs/install.md"},{"url":"https://agentprose.cloud/docs/how-agents-use-your-site","slug":"how-agents-use-your-site","title":"How AI agents use your site","summary":"What WebMCP is, what an AI browser sees on a page with agentprose, which browsers support it, and why your readers see no change.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["webmcp","agents","concepts"],"categories":["Concepts"],"lang":"en","readingMinutes":2,"text":"When an AI agent visits a page today, it gets the whole page: menus, cookie banners, share buttons and somewhere in the middle the article. It has to guess where the content starts, it spends tokens on markup, and it has no reliable way to search your archive. So it misses posts, mixes them up or quotes you wrong. WebMCP is a new browser standard that lets a page hand the agent tools instead. The agentprose script registers a set of read-only tools on every page of your site. An AI browser that supports WebMCP sees them next to the page and can call them directly. What an agent can do - Search your whole archive , ranked, with filters for tag, author and date. - Read a post as clean Markdown , in full or one section at a time, without menus or ads. - Check a quote before citing you, and get a citation with author, date and a link that highlights the exact words. - Find related posts , list your tags and authors, and get your subscribe link. The full list, with every input, is in the WebMCP tools reference. A typical visit A reader asks their AI browser: \"What has this blog written about pricing?\" 1. The agent calls search posts with pricing and gets the matching posts with titles, summaries and links. 2. It calls get section to read just the part of one post it needs. 3. Before quoting, it calls get citation with the sentence. agentprose confirms the text really appears and returns a link that scrolls to it. Each of those calls is one row in your dashboard. Safe by design - E","md":"/docs/how-agents-use-your-site.md"},{"url":"https://agentprose.cloud/docs/webmcp-tools","slug":"webmcp-tools","title":"WebMCP tools reference","summary":"Every tool the agentprose script registers for AI agents, what it returns, and its inputs.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["webmcp","reference","tools"],"categories":["Concepts"],"lang":"en","readingMinutes":2,"text":"These are the tools an AI agent sees on a page with the agentprose script. All built-in tools are read-only and mark post text as untrusted content. Long answers are paged: they stop at max chars (default 20,000) and return a next offset to continue. Tool What it returns --- --- get current article The post being viewed as clean Markdown, with title, authors, dates, tags, reading time and section outline search posts Ranked full-text search across the whole site, with filters list posts Newest or oldest posts, filterable and paginated get post Any post as Markdown by URL or slug, optionally one section get table of contents The heading outline of a post get section One section's Markdown, matched by id or heading text get related posts Posts related by shared tags, categories and topic list taxonomies Tags, categories and authors, with post counts get citation A formatted citation; with quote it checks the text appears and returns a link that highlights it get site info Site name, description, language, feeds, post count and latest post get subscribe link , get comments link , navigate to Opt-in actions you turn on in the script tag Inputs search posts : query (required), limit (1–50, default 10), tag , category , author , since , until (ISO dates). list posts : limit (1–100, default 20), offset , sort ( newest or oldest ), tag , category , author , since , until . get post : url or slug , optional section , max chars , offset . get current article : max chars , offset . get ","md":"/docs/webmcp-tools.md"},{"url":"https://agentprose.cloud/docs/content-discovery","slug":"content-discovery","title":"How we find your content","summary":"How agentprose Cloud builds a site's index from agentprose.json, feeds, sitemaps and llms.txt, what counts as a post, and how to fix \"no posts found\".","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["crawling","indexing","troubleshooting","agentprose.json"],"categories":["Your content"],"lang":"en","readingMinutes":2,"text":"Hosted search and the dashboard's post list come from your site's index . We build it from what your site already publishes, in this order, and stop at the first source that has posts: 1. /agentprose.json , the static index the agentprose CLI writes. It is the richest and cheapest source. 2. Your feed : RSS 2.0, Atom or JSON Feed, found through <link rel=\"alternate\" on your home page or at /feed.xml , /index.xml , /rss.xml , /atom.xml , /feed.json or /feed/ . 3. Your sitemap ( /sitemap.xml , including sitemap indexes) and /llms.txt . We fetch each listed page and keep the ones that are content. Feeds often carry only a summary, so we also fetch those posts' pages for the full text. What counts as a post From a sitemap, a page counts when it has any of: - schema.org JSON-LD of an article type: Article , BlogPosting , TechArticle , NewsArticle , HowTo and similar - <meta property=\"og:type\" content=\"article\" - a single <article element with a <time datetime and some length - a documentation-style path such as /docs/ , /guide/ , /help/ , /blog/ or /kb/ , with real text on it - a link in your llms.txt , with real text on it Tag, category, author, search and pagination pages are skipped. If no page qualifies, we fall back to the longest text pages so the index is never empty for a content site. \"No posts found\" The site's Search index tab shows what we tried and what each source answered. The usual fixes, quickest first: - Add a feed. Most blog platforms have one; make sure it is l","md":"/docs/content-discovery.md"},{"url":"https://agentprose.cloud/docs/hosted-search","slug":"hosted-search","title":"Hosted search","summary":"The hosted search API behind the search_posts tool: request and response, quotas, caching, fallback and semantic search on Team.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["search","api"],"categories":["Your content"],"lang":"en","readingMinutes":1,"text":"With a site key, the search posts tool answers from your site's hosted index instead of downloading your feeds into the agent's browser. It is faster, covers your whole archive and works on sites with thousands of posts. The API It is open to any origin (CORS) and cached for 60 seconds. Ranking is BM25 over title, tags, summary and body. On Team, results are also ranked by meaning (semantic search), not only by matching words. Quotas Free includes 1,000 searches a month, Pro 50,000 and Team 500,000. Over the quota the API answers 429 and the agentprose script quietly falls back to its own search in the browser, so agents still get results. Your dashboard shows when that happens. Keeping it fresh The index is rebuilt by crawling, by CMS webhooks, or by uploading agentprose.json from CI: Keeping search local Add data-cloud-search=\"false\" to the script tag to keep analytics but let the script search in the browser.","md":"/docs/hosted-search.md"},{"url":"https://agentprose.cloud/docs/cms-webhooks","slug":"cms-webhooks","title":"CMS webhooks and re-crawls","summary":"Point WordPress, Ghost, Sanity, Contentful or any CMS at your site's webhook so new and changed posts reach hosted search within seconds.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["webhooks","cms","wordpress","ghost","sanity","contentful"],"categories":["Your content"],"lang":"en","readingMinutes":1,"text":"Every site has a webhook URL, on its Search index page. Point your CMS's \"on publish\" webhook at it and new or changed posts reach hosted search within seconds. Webhooks are on every plan. By platform - Ghost : Settings → Integrations → custom integration → webhooks for Post published , Published post updated , Post unpublished and Post deleted . The post is indexed straight from the webhook; members-only posts are never indexed. - WordPress : a WP Webhooks \"Send Data\" webhook on post create, update and trash, or a few lines on transition post status that post {\"post permalink\": \"...\", \"post status\": \"publish\"} . Published permalinks are re-fetched; any other status removes the post. - Sanity : a GROQ-powered webhook with the projection {\"url\": \"https://your.site/posts/\" + slug.current + \"/\"} . Create and update re-fetch the page, delete removes it. - Contentful : a webhook on Entry publish, unpublish, archive and delete. Entries carry no public URL, so each event queues one debounced re-crawl. - Strapi, Storyblok and others : any POST works (below). Any other system Add \"action\": \"delete\" to remove a post, or send an empty body to re-crawl everything. Static sites Static sites can skip webhooks and upload agentprose.json from CI (see Hosted search). Scheduled re-crawls Each site is also re-crawled on a schedule you set in its settings: daily by default, down to every 6 hours on Pro and every hour on Team. Re-index now on the Search index tab runs one at once.","md":"/docs/cms-webhooks.md"},{"url":"https://agentprose.cloud/docs/analytics","slug":"analytics","title":"Agent analytics","summary":"What the agentprose Cloud dashboard shows about AI agents on your site, exactly what the script sends, and how agents are identified.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["analytics","dashboard","privacy"],"categories":["Analytics"],"lang":"en","readingMinutes":1,"text":"Your dashboard shows every tool call an AI agent makes on your site, within seconds: - Which agents came : ChatGPT Atlas, Gemini in Chrome, Perplexity Comet, Claude, Copilot and more. Agents we can't identify are labelled as such, not guessed. - What they asked : the searches agents ran on your site. - What they couldn't find : searches with no result, ranked by how often they came up. That is your content-gap report, and often your next post. - What they quoted : which posts get cited, and whether the quote matched your text. - Speed and errors : response times and failed calls per tool. Everything is available as CSV and through the API on every plan. What the script sends One small beacon per tool call, with navigator.sendBeacon and no cookies: query is sent only for search tools (at most 200 characters) and can be turned off per site. n is the number of results, found the citation verdict, post the URL an agent asked for. Post content is never sent. What we store The page path without its query string, the agent family derived from the User-Agent (not the string itself), and a daily-rotating hash for unique counts. Raw IP addresses are never stored. See Privacy and data. Limits Free includes 10,000 agent events a month, Pro 250,000 and Team 2,000,000. Over the limit nothing breaks on your site: events above 110% of the plan are not stored, and the dashboard tells you.","md":"/docs/analytics.md"},{"url":"https://agentprose.cloud/docs/api","slug":"api","title":"Management API","summary":"Manage sites, read dashboard numbers, upload an index and queue crawls with a secret API key.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["api","reference"],"categories":["Analytics"],"lang":"en","readingMinutes":1,"text":"Authenticate with Authorization: Bearer sk … . Create keys under API keys in the dashboard; we store only a hash, so copy the key when it is shown. Secret keys belong on servers and in CI, never in a page. Method Path What --- --- --- GET /v1/sites Your sites GET /v1/sites/:id/stats?days=30 The dashboard numbers as JSON PUT /v1/sites/:id/index Replace the index with an agentprose.json body POST /v1/sites/:id/crawl Queue a crawl Examples The public search endpoint is described in Hosted search. The management API is rate-limited to 120 requests a minute per client.","md":"/docs/api.md"},{"url":"https://agentprose.cloud/docs/pricing-and-billing","slug":"pricing-and-billing","title":"Plans, pricing and billing","summary":"What Free, Pro and Team include, how VAT works at checkout, and how to upgrade, change or cancel a plan.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["pricing","billing","faq"],"categories":["Account"],"lang":"en","readingMinutes":1,"text":"Free Pro Team --- --- --- --- Monthly 0 NOK 159 / €14 NOK 690 / €59 Yearly (2 months free) 0 NOK 1,590 / €140 NOK 6,900 / €590 Sites 1 3 20 Agent events / month 10,000 250,000 2,000,000 Hosted searches / month 1,000 50,000 500,000 Posts per site in the index 500 5,000 20,000 Scheduled re-crawl, at most every 24 h 6 h 1 h Retention 30 days 1 year 2 years Team seats 1 3 10 Semantic search – – ✓ CMS webhooks, the index upload API and CSV export are on every plan, Free included. Current prices are always on the pricing page. Questions Do I need a card for Free? No. Free has no time limit. How is VAT handled? Prices exclude VAT. Checkout runs on Stripe, and Link (\"Sold through Link\") is the seller of record: it works out the VAT or sales tax for your country and adds it before you pay, for example 25% MVA in Norway. Your invoices come from Link. NOK or EUR? Choose on the pricing page. The price is set by hand in each currency, not converted. How do I upgrade, switch plans or cancel? Dashboard → Billing . Upgrades start at once. Changes, invoices and cancellation are in the billing portal linked from there. A cancelled plan runs to the end of the period you paid for. What happens if I go over a limit? Nothing breaks on your site. Events above 110% of your plan are not stored, and hosted search answers 429 so the script falls back to its own search. The dashboard shows it, and you can upgrade any time. Promotion codes? Enter them on the checkout page under \"Add promotion code\". Anyt","md":"/docs/pricing-and-billing.md"},{"url":"https://agentprose.cloud/docs/privacy-and-data","slug":"privacy-and-data","title":"Privacy and data","summary":"What agentprose Cloud stores about agent visits, what it never stores, where the data lives, how long it is kept and how to delete it.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["privacy","gdpr","data"],"categories":["Account"],"lang":"en","readingMinutes":1,"text":"agentprose Cloud measures AI agents, not people. It sets no cookies on your site and needs no consent banner of its own. What we store - The tool an agent called, whether it worked and how long it took - The page path, without its query string - For searches: the query (at most 200 characters; you can turn this off per site) and the number of results - The agent family, derived from the User-Agent (not the string itself) - A daily-rotating hash for unique counts What we never store Raw IP addresses, the full User-Agent string, cookies, and the content of your posts. The script never sends post content; hosted search indexes the public pages your site already publishes. Where and how long Data is hosted in Norway East. It is kept for your plan's retention (30 days on Free, 1 year on Pro, 2 years on Team) and deleted with your account. Your role and ours For your site's analytics data you are the controller and agentprose Cloud is the processor. Our data processing agreement covers this. The privacy policy describes the data about you as a customer, and the terms the service. Deleting data Deleting a site removes its events and index. Deleting your account removes everything. Questions: hello@agentprose.cloud.","md":"/docs/privacy-and-data.md"},{"url":"https://agentprose.cloud/docs/open-source","slug":"open-source","title":"Open source and self-hosting","summary":"agentprose Cloud runs on the MIT-licensed agentprose library. Use the library on its own without an account, build a static index with the CLI, or add the Cloud for analytics and hosted search.","published":"2026-10-18T00:00:00.000Z","updated":"2026-10-18T00:00:00.000Z","authors":["agentprose Cloud"],"tags":["open-source","self-hosting","cli"],"categories":["Account"],"lang":"en","readingMinutes":1,"text":"The tools agents use on your site come from agentprose, an open-source library under the MIT license. agentprose Cloud adds analytics and hosted search on top. You can use the library without an account. The library on its own With no configuration it reads your page, your agentprose.json , your feeds, your sitemap and your llms.txt in the visitor's browser, and registers the same WebMCP tools. No backend and no account. You don't get the dashboard or hosted search. Static index with the CLI For large static sites, build the index at deploy time: build writes agentprose.json , one Markdown file per post and, with --llms , an llms.txt . inspect shows what agents can read on a live page and gives it an agent-readiness score. In your own code createAgentTools() lets you pick an adapter (WordPress, Ghost, Sanity, Contentful, Strapi, Storyblok, Enonic), add your own tools and send tool-call events to any analytics. See the library's README. This site uses it too agentprose Cloud runs its own script on these pages and publishes its docs as agentprose.json, an Atom feed and llms.txt. Open any docs page with ?agentprose-inspect to see the tools.","md":"/docs/open-source.md"}]}